Numerous client devices deploying an application may experience crashes of the application. When a client device experiences a crash, a crash “fix” is developed. The crash fix is sent to client devices on which the application is deployed. In an example, the crash fix is integrated into an updated version of the application, and the updated version of the application is sent to client devices having the application. Sending crash fixes or updated versions of the application to client devices may consume a significant amount of network bandwidth and computing power, which may be unnecessarily consumed when less than all client devices may need the crash fixes (e.g., some client devices may not experience the crash, such as due to different hardware and/or software configurations that may not induce the crash).